Subscribe to our newsletter to receive the latest news and events from TWI:

Subscribe >
Skip to content

Some preliminary evaluations of black coating on aluminium AA2219 alloy produced by plasma electrolytic oxidation (PEO) process

   
S. Shrestha (1), A. Merstallinger (2), D. Sickert (3) and B. D. Dunn (4)

(1) TWI Ltd, Cambridge, United Kingdom,
Email: suman.shrestha@twi.co.uk
(2) ARC Seibersdorf research GmbH, Austria,
Email: andreas.merstallinger@arcs.ac.at
(3) Dresden University of Technology, Germany,
Email: Daniel.Sickert@mailbox.tu-dresden.de
(4) European Space Agency, Noordwijk ZH, The Netherlands,
Email: Barrie.Dunn@esa.int

Paper presented at 9th International Symposium on Materials in a Space Environment, 16-20 June 2003, ESTEC, Noordwijk, The Netherlands.

Abstract

This paper describes the results of a study of a black coating produced on aluminium AA2219 alloy using a process that involves creation of a hard ceramic oxide layer on the surface of the alloy by plasma electrolytic oxidation (PEO) known as the 'KERONITE ®' process.

Coating microstructure has been examined and the coating characteristics such as porosity, hardness, adhesion and phase composition were measured. The thermo-optical properties such as solar absorptance ' αs' and normal infrared emittance ' ε n-IR' of the coating were measured in the 'as-prepared' condition and after environmental exposures to humidity, thermal cycling and UV-radiation in vacuum and to thermal shock. Comparison was made with alternative coatings produced using standard black anodising processes.

The study also looked at the cold welding and friction behaviours of the coated alloy in vacuum and in an ambient laboratory environment. Standard spacecraft materials tests were conducted on the coated disc against an AISI 52100 steel ball and also against a coated pin using a pin-on-disc apparatus. Parameters such as friction coefficient and wear depth were measured and the cold welding behaviours were investigated. Test results were compared with the data generated for NiCr plated and anodised coatings. Corrosion performance was assessed using a salt spray exposure test and using an accelerated electrochemical test method. In addition, the study looked at the effect of post coating sealing with a sol-gel solution.

1. Introduction

Coating application by electrolytic methods such as anodising has been extensively used for improving the wear (abrasion) and corrosion resistance of aluminium alloys. The porous nature of anodised coatings allows production of coloured coatings by deposition of organic dyestuffs or metallic pigments. [1] For space applications, additional coating characteristics such as black colour effect to improve thermo-optical properties e.g. solar absorptance, infrared emittance and improved surface characteristics to resist against cold welding are considered necessary. A widely used coating technique to date for space applications is black anodising that has been considered to provide aluminium alloys with suitable thermo-optical properties and acceptable resistance to corrosion, wear and cold welding. However, the conventional black anodising process is considered environmentally not safe and the black colour finish for long-term exposures is not stable. Thus, there exists a strong need to identify new coating processes that can meet the requirements of recent environmental legislation and the continual drive for better coating performance. The 'KERONITE ®' process is one of the relatively new environmentally safe electrolytic coating processes that is applicable to light metals and their alloys in particular Al and Mg.

The 'KERONITE' process involves the application of a modulated voltage to the component in an electrolytic bath agitated using compressed air. The voltage is sufficiently high to create intense plasma due to micro-arc generation at the component surface. This results in oxidation of the component surface (plasma electrolytic oxidation) as well as elemental co-deposition from the electrolyte solution, which creates a hard ceramic oxide layer on the substrate alloy. A typical Keronite coating consists of a porous outer layer and a low porosity main layer that is the bulk of the coating. The top porous layer is often removed by polishing to expose the main layer. The electrolyte is a low concentration alkaline solution of proprietary composition. Similar coating processes also known as micro arc oxidation (MAO) have been reported elsewhere. [2]

2. Experimental approach

2.1 Materials and coating characteristics

An aluminium alloy type AA2219 of composition bal Al, 6.6Cu, 0.3Mn, 0.1Si, 0.01Mg, 0.1Fe, 0.1V and 0.1Zr was selected in this work as a substrate material. Coatings to a thickness of 75-80µm were prepared at Keronite Ltd on disc specimens of various sizes cut from a 6mm rolled AA2219 plate. The disc specimens were ground to 600-grit finish using SiC paper and the edges of the discs were rounded to 2mm radius prior to coating deposition. One face of the coated specimen was polished to a final coating thickness of about 60-70µm after removing a top porous layer. This involved manual polishing of the coating face successively on 360 and 600 grit SiC papers followed by final polishing in 1µm diamond slurry to Ra ~0.04µm. The coating after polishing was rinsed in running water and degreased with acetone. These are referred to 'as-prepared' coatings. In addition, the coatings were subjected to post sealing by dipping in a colourless sol-gel solution followed by subsequent further polishing to remove the top sol-gel layer. This resulted in the exposure of the Keronite surface while the pores remained impregnated and sealed with sol-gel. The latter coatings are referred to 'sealed' coatings.

Coating cross sections were prepared, ready for examination using standard metallographic techniques to a 1µm diamond polishing. A scanning electron microscope (SEM) was used for coating microstructure characterisation. Coating porosity was measured from the polished cross-section using the In2ViewCat-Pro image analyser connected to an optical microscope. Coating hardness was measured from the polished cross section using a DURAMIN Vickers diamond pyramid micro-indenter from Struers Ltd and using a 50g indentation load. Determination of various crystalline phases in the coating was undertaken using the X-ray diffraction (XRD) technique by collecting spectrum from the coating surface using CuK α radiation.

Coating adhesion measurements were undertaken following the guidelines as described in the ASTM standard C633-79 (Re-approved 1993) 'Standard test method for adhesion of cohesive strength of flame sprayed coatings'. The test method was slightly modified from the ASTM C633-79 and consisted of bonding a steel dolly to the coated face of the test specimen (for this only one face was coated) and another identical steel dolly bonded to the rear uncoated face of the test specimen. The test specimen was sandwiched between the two steel dollies using an epoxy film adhesive, type FM73 and heat cured at 120°C. The bonded test specimens were then individually placed in a tensile loading machine (Dartec) with self-aligning devices. Tensile load was increased at 1mm.min -1 and the load at failure recorded. The failure stress was calculated and the nature of failure, whether at the coating/substrate interface, coating cohesion or in the adhesive, was examined. Adhesion measurements were taken also from the coated specimens that had been subjected to 336 hours of salt spray exposure.

The coating peel-off test followed the procedure as described in the European Cooperation for Space Standardization ECSS-Q-70-13A 'Measurement of the peel and pull-off strength of coatings and finishes using pressure sensitive tapes'.[3] This test method is based on the controlled peeling of a pressure sensitive tape from the sample surface (Ø25mm) using a tensile loading machine. The maximum failure stress was recorded and the sample surface as well as the tape was examined to look at particles detached from the coating surface that may have adhered to the tape. The tape used was double sided 3M-pressure sensitive tape type 600 (670g.cm -1). Details of the test procedure are described in ECSS-Q-70-13A. [3] This method was used to look at dust generation from the as-prepared coating surface and the coating that had been exposed to thermal shock.

2.2 Environmental exposure

Humidity

During the humidity test coated samples were exposed for one week to 95% relative humidity (RH) at 50°C. The equipment used was a Heraeus Vötsch humidity chamber.

Thermal cycling in vacuum

The thermal cycling test was performed according to ECSS-Q-70-04A. [4] It comprised 100 cycles of exposure from +100 to -100°C in vacuum. The heating and cooling rate was 10°C·min -1 with the dwell time of 10 minutes at the maximum and minimum temperatures.

UV-exposure in vacuum

The UV-exposure was carried out in a vacuum chamber at a pressure between 10 -6-10 -7mbar and at a temperature of 22-25°C. The coating sample was irradiated using a Hamamatsu deuterium lamp having a wavelength spectrum from 115 - 400nm. The distance of UV-source to the coating surface was about 370mm. The exposure period was 167 hours, which is about 6513 equivalent sun hours (ESH). Details of this procedure are described in the literature. [5]

Thermal shock

Two separate thermal shock exposure tests were undertaken. The first exposure test was conducted at the European Space Agency, which comprised of transferring the coated specimens from a heated oven (+50°C) to a bath of liquid nitrogen (-196°C). This was repeated 10 times with a dwell time of 10 minutes at each temperature. The second exposure test was performed at TWI Ltd. This consisted of alternate immersion of the coated specimens at two temperature extremes: a boiling de-ionised water bath maintained at +100°C and a liquid nitrogen bath at -196°C. The coating specimens were manually transferred and immersed 50 times in each bath with a dwell time of five minutes in each bath. The maximum time during the transfer between the two baths was 10 seconds. The specimens were degreased with alcohol before and after the thermal shock tests followed by air-drying.

Salt spray exposure

The corrosion performance of the coated and uncoated test specimens was examined by exposing the specimens to a salt spray environment following the guidelines described in the ASTM standard B117-97 up to 2000 hours of exposure period. The testing comprised of exposing the coating surface up to 2000 hours to a 5wt% NaCl solution (at 35°C) atomised to create a fog within an enclosed chamber. Changes to the coating surface were recorded following periodic observations at 24, 336, 1000 and 2000 hours. The surface quality was given a rating number in accordance with ASTM D1654-92 'Evaluation of painted or coated specimens subjected to corrosive environments - procedure B'. Corrosion resistance to a minimum of 336-hour test duration is defined in the Aerospace Materials Specification AMS 2470J (R) 'Anodic treatment of aluminium alloys - chromic acid process' as acceptance criteria. [6]

Electrochemical corrosion

The electrochemical corrosion behaviour was studied using an accelerated potentiodynamic test (anodic polarisation) in a de-aerated 3.5% NaCl solution of pH 8 at 25°C using a standard three-electrode test method in an Avesta cell. The coated surface was exposed to the electrolyte and the rest potential 'E corr' (also known as the free corrosion potential) was allowed to stabilise for one hour prior to anodic polarisation. The potential value was measured using a reference saturated calomel electrode (SCE) and a plot of the current density as a function to the polarisation potential was recorded using a platinum auxiliary electrode. More detail of the test procedure can be found elsewhere. [7] The collected polarisation plots were used to compare the electrochemical corrosion behaviour of the coatings with uncoated alloys and the integrity of coatings before and after a thermal shock exposure.

2.3 Measurement of thermo-optical properties

Solar absorptance ' α S'

The absolute reflectivity spectrum of the specimen was measured using a Cary-500 UV-VIS-NIR photo-spectrometer with an integrating sphere accessory and corrected with the help of a reference standard. The solar reflection coefficient was calculated by integrating the measured absolute reflectivity spectra over the solar energy spectrum [8] within the considered wavelength range from 250 - 2500nm. This wavelength band comprises 97% of the energy radiated by the sun. The result of this integration was divided by the total solar energy thus providing the ratio of reflected to incident solar energy (R S). This can be translated into the solar absorptance ' α S' by subtracting it from unity for completely opaque samples, i.e. if the transmittance equals zero. More details on this can be obtained in the references. [9,10]

Normal infrared emittance ' ε n-IR'

The normal infrared reflectivity (R n-IR) of the coating surface was measured using a Gier-Dunkle DB100 infrared reflectometer. According to Kirchhoff's radiation law the infrared emittance of a material equals its infrared absorption when maintained at the same temperature. Furthermore the amount of radiation, not reflected by non-transparent materials, equals the absorbed portion of radiation. The measurement device consisted mainly of an internal thermal source that emits infrared radiation towards the coating surface and a detector for the amount of radiation reflected. Applying the above-mentioned relations this value is translated into the normal infrared emittance ' ε n-IR' by subtracting it from unity. More detailed description of the physical basics can be found in the literatures. [9,10]

2.4 Cold welding and friction behaviour

For the cyclic impact adhesion test, a contact was made between a vertically moving pin. The pin was either AISI 52100 bearing steel of nominal composition bal Fe, 1C, 0.3Si, 0.4Mn, 1.6Cr, 0.3Ni, 0.3Cu or coated aluminium alloy having a radiused contact area against a fixed flat disc (coated or uncoated aluminium alloy) for about 5000 cycles. Each cycle consisted of three steps: i) an impact loading; ii) a static load held for 10 seconds at a vacuum pressure of less than 5.10 -8mbar; and iii) unloading with measurement of the separation force, i.e. adhesion force.

For the cyclic fretting adhesion test, a contact (without impact) was made followed by a static load, which is held for 10 seconds at a vacuum pressure of less than 5.10 -7mbar. During this time fretting is applied: the pin was moved with a sinusoidal frequency at 210Hz and a stroke of 50µm in each cycle for a total of 5000 cycles. After stopping fretting, the force required to separate the pin from the disc was measured. Tests were conducted on the coated and uncoated test specimens. The uncoated test specimens were fine ground to surface roughness of Ra < 0.03µm. The coated face was in the polished condition to Ra ~ 0.04µm.

Sliding wear tests were conducted using a vacuum pin-on-disc tribometer manufactured by CSEM. The pin (ball) was allowed to slide on the disc surface in unidirectional movement at a sliding speed of 0.1ms -1 and a sliding distance of 1000m. Vacuum pressure was maintained at less than 10 -6mbar and the temperature at 25°C. The ball diameter was 7mm and the load was 10N. More details on the testing procedures are given in the literatures. [11,12] In addition, sliding wear tests were undertaken also in an ambient laboratory environment (22±2°C, relative humidity 35-45%) with a test load of 10N and sliding speed of 0.1ms -1 against a steel ball (AISI 52100) of diameter 10mm for a 1000m sliding distance. This followed the guidelines described in ASTM G99-95a 'Standard test method for wear testing with a pin-on-disc apparatus'. Friction coefficients and wear track depths where possible were measured.

3. Results and discussions

3.1 Coating characterisation prior to exposure

A backscattered SEM image of a sealed coating cross section is shown in Fig.1, which shows a relatively dense coating. This image is of the main coating layer after removing the top porous layer (not shown), which also reveals a microstructure comprising of interlocking grains about 20-30µm in size with occasional very small-sized pores within the grains and a few very fine microcracks. The coating porosity was measured at about <3%.

spssjun03f1.jpg

Fig.1. SEM backscattered image of the sealed Keronite coating cross-section on AA2219 alloy

A backscattered SEM image of a sealed coating surface is shown in Fig.2. The image in Fig.2 shows the presence of porosity on the polished surface. The image in Fig.2 also shows the presence of a lighter contrast phase that lies at the boundaries between the interlocking grains. The lighter contrast in the backscattered SEM image is indicative of this phase being of a higher atomic number element.

spssjun03f2.jpg

Fig.2. SEM backscattered image of a polished surface of the sealed Keronite coating on AA2219 alloy
1 - Al 2O 3; 2 - Cu containing phase; 3 - Sealant.

Spot EDX analyses taken from the areas in Fig.2 showed that area '1' comprised of predominantly 'Al' and 'O' peaks suggesting it to be aluminium oxide. The lighter contrast phase '2' at the grain boundaries were found to be highly rich in 'Cu' and such could have come from the copper precipitates of the substrate alloy that do not participate in the oxidation process. The darker phase '3' was found to be rich in 'Si' content and was believed to be impregnated sealant. A cross section through the specimen edge is shown in Fig.3. The coating appeared well retained over the edge with no visible cracking over the tight radius.

spssjun03f3.jpg

Fig.3. SEM image showing good and uniform Keronite coating retention at the substrate edge

The X-ray diffraction pattern collected from the coating surface is presented in Fig.4, which suggested that the Keronite coating comprised predominately α and some γ and δ-Al 2O 3 crystalline phases. This has resulted in a very high coating hardness at an average value of about 1369HV with the values ranging between 1225-1524HV. The hardness measured on the substrate alloy was only about 140HV. This is indicative of the coating hardness being about 10 times higher than the substrate alloy.

spssjun03f4.gif

Fig.4. XRD patterns of the Keronite coating

The mean tensile stress to failure during the pull-off test from four coating specimens was recorded at greater than 28MPa with the failure occurring mainly within the adhesive. It is expected the coating would have adhesion in excess of 30MPa. The results of the peel-off test for the sealed coatings showed the primary mode of failure was adhesive tape separation from the coating without any particle removal.

3.2 Coating characterisation after environmental exposures

Examination of surfaces of the sealed coatings after the thermal shock exposure did not show any visible coating degradation. However, there was a thin band of the coating removal only at the edge of the disc specimen thus exposing the underlying substrate. This would have been expected due to stress concentrations at the specimen edge.

The values of adhesion measured for the coatings after the thermal shock were in excess of 25MPa and the mode of failure occurring primarily within the adhesive. Adhesion of the coatings after 336-hour salt spray exposure was measured and was in excess of 30MPa with failure in the adhesive. The coating peel-off test after the thermal shock resulted in the tape separation without any debris being seen on the peeled adhesive tape. The average coating hardness measured from the cross section of the specimen that had been exposed in a salt spray environment for 2000 hours displayed about 1500HV. These results demonstrate that the coating did not show any loss of its integrity and retained near original microstructure with good adhesion and high hardness after the environmental exposures.

3.3 Corrosion behaviour

Exposure of the as-prepared coating (unsealed) to the salt spray environment displayed severe corrosion product seen on the coating surface just after 120 hours. This corrosion attack was in the form of blisters and extended pits with pit diameters about 1mm were observed on the coating surface. Surface appearance of the 'as-prepared' coating (unsealed) after 120 hours of salt spray is shown in Fig.5. Such would have been expected from the as-prepared (unsealed) coating having porosity and interconnected path for the corrosive media to penetrate into the aluminium substrate.

spssjun03f5.jpg

Fig.5. Surface of the as-prepared (unsealed) Keronite coating on AA2219 alloy after 120 hours of salt spray

Further salt spray exposure tests were undertaken for the coated specimens sealed with sol-gel, which displayed significant improvement in the corrosion resistance afforded by the sealed coating to the underlying aluminium substrate. The sealed coating surface displayed significant protection in the salt spray environment with no visible corrosion observed and was rated '10' (the highest rating) after 336 hours of exposure. On further exposure up to 2000 hours, the surface of the sealed coating displayed only a very slight corrosion attack on the exposed surface and was still rated '9' at 2000 hours. In contrast to this, the uncoated surface of the AA2219 alloy exhibited very poor performance in the salt spray environment when examined at 24 hours and was rated as zero. Surface appearances of the uncoated alloy and the sealed coating after 2000 hours of salt spray are shown in Fig.6. A coating section though the sealed coating after exposure to 2000 hours in a salt spray environment is shown in Fig.7, which displays no visible deterioration of the coating or the substrate underneath.

Fig.6. Surface of: 6a) uncoated
Fig.6. Surface of: 6a) uncoated
6b) sealed Keronite coating on AA2219 after 2000 hours of salt spray
6b) sealed Keronite coating on AA2219 after 2000 hours of salt spray
Fig.7. Cross-section of the sealed Keronite coating on AA2219 alloy after 2000 hours of salt spray
Fig.7. Cross-section of the sealed Keronite coating on AA2219 alloy after 2000 hours of salt spray

Anodic polarisation plots from the electrochemical tests are shown in Fig.8. The uncoated 2219 alloy displayed an immediate rapid increase of the current density on slightly increasing the potential from its rest potential 'E corr'. This immediate rise of the anodic current density over a small positive potential range from the rest potential is indicative of very rapid corrosion occurring at the surface of the uncoated alloy. The coated alloy displayed a large potential range (passivity) positive from the E corr, where the current density was stable and low at 0.2µA.cm -2, indicative of very little corrosion occurring on the coating/substrate system and a good level of barrier protection afforded to the underlying substrate. Higher currents would be expected if the salt solution penetrated into the underlying substrate via cracks and pores.

Fig.8. Anodic polarisation plots (forward scans) from the uncoated AA2219 alloy, sealed Keronite coating before and after the thermal shock exposure
Fig.8. Anodic polarisation plots (forward scans) from the uncoated AA2219 alloy, sealed Keronite coating before and after the thermal shock exposure

Anodic polarisation of the coating after exposure to thermal shock displayed the stable current density region being shifted to about 2µA.cm -2. This was taken to indicate that possible fine cracks might have resulted during the thermal shock, which possibly had led to a slightly higher corrosion rate from the substrate. However, this value was still low and a large potential region at stable current density < 5µA.cm -2 would mean the coating was still effective as a barrier to minimise the passage of the corrosive electrolyte to the underlying aluminium substrate.

3.4 Thermo-optical properties

The mean solar absorptance ' α S' and infrared emittance ' ε n-IR' values measured from 12 as-prepared Keronite coatings (unsealed) prior to environmental exposures are given in Table 1. The mean ' α S' about 0.89 (±0.01) and ' ε n-IR' ranging between 0.64-0.76 with a mean value of 0.72 was measured.

The ' α S' and ' ε n-IR' values after exposures to high humidity, thermal cycling in vacuum, humidity followed by thermal cycling in vacuum, UV radiation and thermal shock between temperature extremes of +50 and -196°C are given in Table 1. The data in Table 1 showed a good stability of the thermo-optical properties retained by the Keronite coating after the environmental exposures. Visual examination of the coating surfaces did not show any coating degradation after environmental exposures to humidity, thermal cycling and UV radiation. The coating subjected to thermal shock had some coating spalling only from the specimen edge and believed to be due to stress concentrations. No coating damage was observed on the main exposed coating surface after the thermal shock test. The thermo-optical properties after the thermal shock exposure were stable and similar to the coatings prior to the exposure.

Table 1. Solar absorptance ' α S' and infrared emittance ' ε n-IR' of the coatings

Coatings and environmentα Sε n-IRα S/ ε n-IRRef.
Keronite coatings
(Before and after environmental exposure)
Before exposure 0.89 0.72 1.2


-
After humidity 0.89 0.72 1.2
After thermal cycling 0.89 0.71 1.2
After humidity + thermal cycling 0.89 0.73 1.2
After UV exposure 0.89 0.74 1.2
After thermal shock 0.88 0.75 1.2
Benchmark coatings
ESA PSS 01-703 0.96 0.93 1.0

5
Colinal 3100 0.89 0.93 1.0

The data in Table 1 suggests that the solar absorptance of the black Keronite coating can be similar to Colinal 3100, however, the normal infrared emittance is about 20% lower compared to the coatings produced with black anodising processes. The PSS-01-703 coating and Colinal 3100 are based upon the colouring of anodized surface layers. Colouring is done either by depositing cobalt or nickel sulphides or by electrolytically depositing tin. These latter two processes are considered to have environmental concerns.

3.5 Cold welding behaviour

The data from the impact test in Fig.9 shows that the adhesion measured after the impact tests for the Keronite coatings against itself and against a standard bearing steel (AISI 52100) are very low and are similar to anodised and NiCr-plated coatings. However, there were some differences in the damage mechanisms between the Keronite and anodised coatings. The Keronite coating did not display any damage to the coating surface under impact, whereas, the anodised coating displayed breaking of the coating surface under impact resulting in the formation of loose debris on the impacted surface. The aluminium alloy type AA7075 without a coating demonstrated a very high adhesion value after the impact tests. Tests were not undertaken on the uncoated AA2219 alloy.

Fig.9. Adhesion force measured during impact tests
Fig.9. Adhesion force measured during impact tests
Fig.10. Adhesion force measured during fretting tests
Fig.10. Adhesion force measured during fretting tests

The adhesion values measured after the fretting tests are presented in Fig.10. The data in Fig.10 shows that adhesion values of the Keronite coated AA2219 against a 52100 steel ball and NiCr plated AA7075 against an anodised AA7075 are similar. An aluminium alloy without a coating resulted in extremely high adhesion after fretting. The Keronite coating surface against itself displayed a higher adhesion than against a steel surface by about three times.

Despite low adhesion values displayed by the Keronite and hard anodised coatings against steel surfaces, the Keronite coating displayed superiority over the anodised coasting. The Keronite coating showed no sign of coating damage, whereas, the anodised coating displayed extensive cracking and chipping of the coating surface under fretting environment. The differences in the damage mechanisms between the Keronite and anodised coating surfaces are shown in Fig.11.

Fig.11. Surface after fretting test of: 11a) Keronite coating (no visible damage)
Fig.11. Surface after fretting test of: 11a) Keronite coating (no visible damage)
11b) anodised coating (showing extensive cracks)
11b) anodised coating (showing extensive cracks)

3.6 Friction and wear behaviour

The results of the friction wear tests are given in Table 2. The data in Table 2 show mean friction coefficient and disc wear track depth after the sliding wear tests conducted in vacuum and in an ambient laboratory environment.

Table 2. Pin-on-disc sliding wear test data.

Disc/pinEnvironmentMean friction coefficient
(µ)
Disc wear track depth
(µm)
AA2219/AISI 52100 Vacuum 0.3 18
As-prepared Keronite on AA2219/AISI 52100 0.6 -
Keronite AA2219/Keronite AA2219 0.5 -
 
AA2219/AISI 52100 Laboratory ambient temperature (22±2°C) 0.3 26
Sealed Keronite on AA2219/AISI 52100 0.5 (-0.8)
Thermal shocked Keronite AA2219/AISI52100 0.5 (-1.0)

The mean coefficient between an AA2219 disc and 52100 steel ball in vacuum of 0.3 was about twice less than that was observed with the Keronite/steel or Keronite/Keronite surfaces. Similar friction values were measured in the laboratory environment at ambient temperature. In all cases, the Keronite coating was intact without any visible coating damage when in sliding contact with steel. The only damage seen was either on an uncoated AA2219 disc against a steel ball or on a steel ball sliding against the Keronite coated disc. The values in brackets in Table 2 are negative wear i.e. depth loss on a counter steel ball. An exposure to the thermal shock did not appear to reduce the coating sliding wear resistance against the steel ball. However, the Keronite coating sliding against itself displayed some evidence of coating damage on both the coated disc and coated ball surfaces. These are not shown.

Fig.12. Topography of: 12a) Keronite coating surface on AA2219 alloy (shows small particles pulled out)
Fig.12. Topography of: 12a) Keronite coating surface on AA2219 alloy (shows small particles pulled out)
12b) AA2219 alloy (uncoated) after sliding wear test against AISI 52100 steel ball in vacuum (shows deep wear track)
12b) AA2219 alloy (uncoated) after sliding wear test against AISI 52100 steel ball in vacuum (shows deep wear track)

The surface topographies of the Keronite coated AA2219 and uncoated AA2219 discs after sliding wear tests against the AISI 52100 steel ball in vacuum are shown in Fig.12. The Keronite coated surface did not show any sign of visible coating damage, except for a few particle pullouts. The uncoated disc displayed a deep wear track with severe grooving effect on the disc of about 30µm deep and material being displaced from the groove resulting in the formation of raised lips at the edge of the groove.

4. Conclusions

  • The Keronite ® process is capable of depositing a uniform coating at edges and tight corners with a typical coating thickness about 60-70µm. The coating on the AA2219 alloy after polishing is black in colour with aporosity level <3%. The Keronite coating on AA2219 alloy comprises of predominantly crystalline α-Al 2O 3 that results in high coating hardness of about 1300HV. The coating adhesion can be in excess of 30MPa. The individual oxidised particles (grains) in the coating retain good cohesion as shown by the peel-offtest.
  • The corrosion resistance of the as-prepared Keronite coating is poor and a life of less than 120 hours is expected in a salt spray environment. However, this poor corrosion performance can be overcome by post coating sealing. TheKeronite coating sealed with sol-gel can provide extremely good resistance in excess of 2000 hours without a sign of any visible corrosion attack.
  • The Keronite coating on the AA2219 alloy can retain its good integrity and original microstructures with good adhesion, good cohesion and high hardness after the environmental exposures such as salt spray and thermal shock.
  • The solar absorptance to infrared emittance ratio of the Keronite coating is 1.2, suggests that it can replace the PSS-01-703 black anodised coating. The thermo-optical properties of the Keronite coating on AA2219 alloy will remainstable and not be affected by environmental exposures.
  • The Keronite coating on AA2219 can prevent cold welding under impact and fretting against AISI 52100 steel in vacuum environment. The Keronite coating compared to anodised coating can have improved resistance to impacts andfretting, as shown by no visible damage to the Keronite coating after the tests.
  • The Keronite coating sliding against AISI 52100 steel will result in the wear of mainly the steel surface. The coating however, against itself in sliding wear condition will result in higher damage to the coating surfaces.
  • Friction coefficients of the Keronite coating against itself and steel are high at about >0.5.

 

5. Acknowledgements

The authors would like to thank Dr E Semerad and Mr W Costin of Austrian Research Centre Seibersdorf for undertaking fretting, impact and wear tests. Thanks are also due to several TWI staff for undertaking various tests and Dr A Sturgeon for helpful comments to this paper. The authors would also like to acknowledge DePuy International, European Space Agency, Goodrich Actuation Systems and Keronite Ltd. for the permission granted to use data on AA2219 alloy generated in the TWI Group Sponsored Project funded by them.

6. References

  1. Wernick S. Pinner R. and Sheasby P.G., The surface treatment and finishing of aluminium and its alloys, 5 th ed., Vol. 1, ASM International and Finishing Publ. Ltd., 1987.
  2. Dearnley P.A. et al., The sliding wear resistance and frictional characteristics of surface modified aluminium alloys under extreme pressure, WEAR, Vol. 225-229, p.127-134, 1999.
  3. ECSS-Q-70-13A: Measurement of the peel and pull-off strength of coatings, Publ. ESA, October 1999.
  4. ECSS-Q-70-04A: Thermal cycling tests for the screening of space materials and processes, Publ. ESA, October 1999.
  5. Sickert D., Evaluation of the Keronite coating on AA2219-T87 and AA2024-T6 for thermal control in space applications, ESTEC Metallurgy Report No. 3297, April 2002.
  6. AMS 2470J: Aerospace Material Specification (R), Anodic treatment of aluminium alloys - chromic acid process, revised November 1995.
  7. Shrestha S., et al. Improved corrosion performance of AZ91D magnesium alloy coated with the Keronite ® process, Magnesium Technology 2002, TMS Publ., p.283-287.
  8. ASTM E490-73a: NASA/ASTM Reference Solar Spectrum.
  9. ESA PSS-01-709: Measurement of thermo-optical properties of thermal control materials, Publ. ESA, Issue 1, July 1984.
  10. TOS-QMC WI-TO-02: Operation manual for the calculation of the solar absorptance aS based on reflectance measurements used at ESTEC TOS-QMC, ESTEC internal document.
  11. Merstallinger A. and Semerad E., Test method to evaluate cold welding under static and impact loading, Austrian Research Centre Seibersdorf, Issue 2, 1999.
  12. Merstallinger A., et al. Assessment of Keronite for cold welding and friction, ESTEC Metallurgy Report No. 3522, October 2002.

For more information please email:


contactus@twi.co.uk